An Adapted Retinex Algorithm with Complexity Optimization for Mobile Phone Medical Image Enhancement

Grigoryan, Artyom M.; Agaian, Sos S.

doi:10.1117/3.2204748.ch5

Book: Electronic Imaging Applications in Mobile Healthcare

Editor(s): Sos S. Agaian; Jinshan Tang; Jindong Tan

Author(s): Artyom Grigoryan, Sos Agaian

Published: 2016

https://doi.org/10.1117/3.2204748.ch5

Abstract

The recent advances in image enhancement research have important implications for many image processing applications, including medical imaging that requires improving a characteristic or quality of an image and supporting human perception. Recently, several image enhancement algorithms and many applicable systems have been exploited. Research in image enhancement covers a wide variety of topics that use algorithms based on the human visual system, e.g., histograms with hue preservation, JPEG-based enhancement for the visually impaired, and histogram modification techniques. Two major classifications of the existing image enhancement methods can be considered. The first includes methods that process images in the spatial domain, such as the well-known histogram equalization, general or weighted thresholding, and the retinex method. The retinex method was originally proposed by Edward Land as a model of human perception of lightness and color that seeks to give computers a method of making color images more true to the color perceived by human eyes under varying illuminations. Under varying types of illumination, the colors of objects can be misinterpreted or simply lost in digital images, whereas eyes, having a much more complex visual system, can perceive a given color under any color of illumination.